Colquhoun.AI
← All guides

Research Guides

Useful Filters, Unnecessary URLs: A Faceted Navigation Review

· Colquhoun.shop editorial

A visitor selects a topic, region and date range. The results look helpful. Behind that simple interaction, the website may generate dozens of addresses that show almost the same collection.

The research question is not “Are filters bad for SEO?” It is “Which filtered views deserve a discoverable address, and which are just temporary ways to browse?” Answer that before prescribing technical changes.

This worksheet is for an owner or reviewer examining an existing directory, catalogue or resource library. It does not require an exhaustive crawl, and it does not assume that Colquhoun.shop performs these tests automatically.

Start with a small, deliberate sample

Choose one collection you understand. Record its unfiltered URL and five ordinary browsing journeys: one filter, two filters, a different sort order, removal of a filter, and a combination with no matches. Keep the test modest, especially on a large or slow website.

For each journey, save the starting address, controls used, final address, visible result count and a few representative item names. Include a timestamp. A screenshot is useful context, but the address and observed response are the reproducible evidence.

Do not start by generating every possible combination. Your first task is to discover patterns, not demonstrate how many requests your crawler can make.

Separate three kinds of views

Classify the examples using this editorial distinction:

These are review labels, not Google classifications. A filtered page can move between them as the collection changes. Record the reasoning rather than declaring that every URL with a question mark should disappear.

For a destination candidate, ask a colleague to explain its value without naming the filters. “It helps local organisations find evening training” is a stronger editorial reason than “It has three selected checkboxes.” Neither statement alone establishes search demand.

Check whether different journeys create the same destination

Repeat one valid combination in a different selection order. Compare the resulting addresses and actual items. Does removing a filter leave an obsolete parameter? Does changing the sort order change the collection or merely rearrange it?

Google's URL guidance warns that excessive parameter combinations can create many similar addresses. It recommends conventional query syntax, including = for key-value pairs and & between parameters.

Write down the observed duplication before proposing consolidation. Never remove a parameter because its name looks unimportant: it might control language, availability or another meaningful difference.

Make the crawl decision explicit

Google's faceted-navigation guidance distinguishes views intended for potential indexing from filters that need not be crawled. For crawlable filters, it recommends consistent ordering, no duplicate filters, and appropriate 404 responses for empty or nonsensical combinations. Canonical links and nofollow are less effective long-term crawl controls than the approaches described in that guidance.

Treat implementation as a separate, reviewed change. Do not paste a broad robots rule into production based on this worksheet. A pattern that catches a sorting parameter might also block valuable pages. Review representative matches and exceptions with the person responsible for the site's routing and indexing policy.

Copy this decision record

Field What to record
Collection Unfiltered URL and business purpose
Journey Exact filter and sorting actions
Result Final URL, count, sample items and timestamp
Comparison Another URL showing the same or different collection
Intended role Destination, browsing preference, or invalid state
Evidence What supports that role; what remains unknown
Proposed change Specific behaviour, not just “fix SEO”
Safety check Valuable URLs and user journeys that must remain intact
Acceptance test Expected response and visible result for each sample
Owner Decision maker and review date

Example: a resource library

Imagine a fictional library of training courses. “Evening courses in Leeds” has enough current entries to be useful. Alphabetical and newest-first versions contain identical courses. A contradictory combination produces an empty list.

The review should create three separate decisions: whether the location-and-time collection deserves a search destination, how ordering variants should be handled, and what response an impossible combination should return. It should not recommend one blanket treatment for all three.

After an approved change, repeat the same journeys. Check that a visitor can still filter, clear selections, follow a course link and return to the collection. Record unexpected changes before expanding the rollout.

Measure the problem before escalating it

Google's crawl-budget guide primarily addresses large or rapidly changing sites. It also makes clear that crawling is not a guarantee of indexing. A small site should not infer a crawl-budget crisis from a handful of parameter URLs.

Where authorised evidence is available, compare crawler requests, important-page discovery and server behaviour over comparable periods. Keep an indexing observation separate from a crawling observation. Without those records, report the narrower finding: a reproducible URL pattern and a proposed test, not a proven cause of lost traffic.

The useful deliverable is a defensible decision about each kind of view. Fewer URLs are not automatically better; clearer purpose and verified behaviour are.

This guide describes the built-in Web Reader and Keyword Density behavior verified for this publication. Available data depends on the page response; these tools do not measure visits or search rankings.